On 31.01.2012, at 21:50, Joop Boonen wrote:
On Tue, January 31, 2012 3:27 pm, Alexander Graf wrote:
On 31.01.2012, at 15:00, Andreas Färber wrote:
Am 31.01.2012 14:44, schrieb Alexander Graf:
On 31.01.2012, at 14:35, Adrian Schröter wrote:
Am Dienstag, 31. Januar 2012, 14:24:28 schrieb Joop Boonen:
On Tue, January 31, 2012 1:58 pm, Alexander Graf wrote: > On 31.01.2012, at 13:54, Joop Boonen wrote: >> Hi all, >> >> I want to put the Toshiba AC100 kernel (kernel-tegra-ac100) in >> Factory.
This is a patched chomeOS kernel for the AC100 : https://gitorious.org/~marvin24/ac100/marvin24s-kernel
With our goal to make the arm port an official openSUSE port, you should also try to get the patches part our official kernel. Or even better in upstream kernel.
Upstream please, maintaining dozens of patches in our tree doesn't help anyone. It just makes rebasing to new kernels harder.
Easier said than done. Despite the weird name, this is the ac100 kernel used by Ubuntu for their official Oneiric ac100 release.
https://wiki.ubuntu.com/ARM/TEGRA/AC100#Kernel_development
This whole kernel situation on ARM is a serious nightmare ...
Yes. Maybe we should switch to a Linaro kernel. ;)
Every time we derive from the actual upstream kernel, we (as in the openSUSE arm community) have to maintain _everything_ ourselves.
Got a kernel bug? Our problem. Don't support new kernel interfaces? Our problem. Don't feel like rebasing the ~300 SUSE patches on top of the next Linaro kernel version? Our problem. Security bugs? Our problem.
I would much rather the above be the openSUSE kernel folks' problem rather than ours.
It looks like Greg Kroah-Hartman is solving this. http://www.h-online.com/open/news/item/Android-drivers-to-be-included-in-Lin...
That's just for Android specific bits, not the hardware support.
We already have enough on our hands to just get all of the distro built properly.
I agree, I think we should focus mostly on the packages first.
That's why I would prefer to first use this type of kernels. This is what other distro's also do (debian, gentoo, ubuntu, fedora etc), for now, as far as I can see.
Sure, and worst case we can even link them into openSUSE:Factory:ARM for the time being. But in the long run, I'd definitely prefer to have things base on the upstream version.
Using non-upstream kernels will definitely not get us anywhere close to an openSUSE 12.2 with official ARM support. At least not for more than the 1 or 2 devices that actually work there.
I wonder how the roadmap will look like? I think that we won't have a fully working/stable 12.2 release. But I think we should freeze openSUSE:Factory:ARM as openSUSE:12.2:Tumbleweed i.e. openSUSE:Tumbleweed. Because it's very difficult to fix all packages with the continuously breaking openSUSE:Factoy:ARM packages a package fixed soke time ago doesn't build currently anymore. This is worst when an or a lot of other packages depend on this package to build.
I think we should fork off when 12.2 forks off as well and try to hammer out a working 12.2 release that runs on at least a few platforms. We should still keep the Factory tree alive and our main focus to develop stuff on, but we would have something we can hand to people without being afraid that it breaks any minute (or takes another week to rebuild). Alex -- To unsubscribe, e-mail: opensuse-arm+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-arm+owner@opensuse.org